General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: May 27, 2018

Application of the Aero-Hydro-Elastic Model, HAWC2-WAMIT, to Offshore Data fromFloating Power Plants Hybrid Wind- and Wave-Energy Test Platform, P37

Bellew, Sarah; Yde, Anders; Verelst, David Robert

Published in:Proceedings of the 5th International Conference on Ocean Energy (ICOE)

Publication date:2014

Document VersionPeer reviewed version

Link back to DTU Orbit

Citation (APA):Bellew, S., Yde, A., & Verelst, D. R. (2014). Application of the Aero-Hydro-Elastic Model, HAWC2-WAMIT, toOffshore Data from Floating Power Plants Hybrid Wind- and Wave-Energy Test Platform, P37. In Proceedings ofthe 5th International Conference on Ocean Energy (ICOE) Marine Renewables Canada.

Application of the Aero-Hydro-Elastic Model, HAWC2-WAMIT,

to Offshore Data from Floating Power Plants Hybrid Wind- and

Wave-Energy Test Platform, P37

Sarah Bellew, Floating Power Plant Anders Yde, DTUDavid Robert Verelst, DTU

sb@floatingpowerplant.com; anyd@dtu.dk; dave@dtu.dk

The 5th International Conference on Ocean Energy (ICOE),Halifax, Canada

4th November 2014

Abstract

As the depths of sites consented for offshore wind increases, the need to develop floatingfoundations for wind turbines increases, as fixed foundations are only economically viable upto approximately 50 m water depth [2]. Key to developing the floating wind turbine industryis the development of accurate numerical models, which can combine the aerodynamic,hydrodynamic, structural flexibility and mooring components. Very little offshore dataexists, however, in order to validate these numerical models.

Floating Power Plant are the developers of a floating, hybrid wind- and wave-energydevice. The device uses the pitching wave energy devices, not only to increase and smooththe power output from the platform, but also to take the energy from the waves in acontrolled manner, resulting in a stable platform for the wind turbine and a safe harbourfor O&M. They are currently developing the final design for their first full-scale prototype,the P80, which has a width of 80 m. As part of the development, Floating Power Planthave completed 4 offshore test-phases (totalling over 2 years offshore operation) on a 37 mwide scaled test device, the P37.

This paper focuses on the comparison of one of the leading numerical models for floatingwind turbines, developed by DTU Wind Energy, to the offshore data from P37. The nu-merical model couples DTU’s own aeroelastic code, HAWC2, with a special external systemthat reads the output files generated directly by the commercial wave analysis software,WAMIT.

1 Introduction

Areas are being consented for offshore wind farms in increasingly deeper water due to the largeavailable area out of sight from the coast, the better energy resources further offshore, and theneed for wind farms in areas where the depth increases rapidly moving away from the coast.Moving further into deeper waters requires the use of floating foundations, as traditional fixedwind turbine foundations are only economically feasible up to approximately 50 m water depth.There is increasing interest in the offshore renewables industry, not only in providing floating

1

foundations for wind turbines, but in combining technologies. One example of this is theEuropean Union Seventh Framework Programme H2Ocean project (www.h2ocean-project.eu),which develops a proof-of-concept design for a new fully integrated multi-purpose platform tomake the most of the sea resources, fostering activities such as renewable power harvesting+ aquaculture + environmental monitoring, and to assess the impact of such platform atenvironmental and economical levels. Floating Power Plant (a partner within the H2Oceanproject) develop a floating wind and wave energy device, and have already for operated a 37m wide, grid-connected test platform for over 2 years, and are currently working towards theirfirst commercial device, P80, which will be 80 m wide.

Key to developing the floating wind turbine industry is the development of accurate nu-merical models, which can combine the aerodynamic, hydrodynamic, structural flexibility andmooring components. DTU Wind Energy Department have been developing a numerical modelthat couples their own aeroelastic code, HAWC2 with a special external system that reads theoutput files generated directly by WAMIT and generates a system with the same response.The HAWC2 model has been extended to account for the wake effects within arrays of windturbines using the Dynamic Wake Meander Model. The combined model will be referred to asHAWC2-WAMIT from herein.

There is currently very little available empirical data from floating wind turbines with whichto validate the numerical models. This paper focuses on the validation and comparison of bothnumerical models when applied to the offshore data acquired by Floating Power Plants testplatform, P37. The platform is a semi-submersible structure which has a water-plane areasimilar to a cross geometry. The cross-geometry is 37 m wide with a hull through the centrewhich is attached to a mooring-buoy fore of the cross-point. The platform is able to rotate (vane)360 degrees around the turret mooring system. There are 10 pitching wave energy convertersalong the width of the platform, whose motion generates power, stabilises the platform andcauses the platform to rotate around the mooring buoy to face the incident waves. On top ofthe platform, there are three Gaia 11 kW wind turbines, one on either side and one on the aft ofthe hull. The wind turbines are two bladed in a downwind configuration with teeter hinge, fixedpitch stall control and a passive yaw system that aligns the single rotor towards the incomingwind.

The P37 platform has undergone four complete offshore test-phases, during which the mo-tions and accelerations of the platform, each of the floaters and the wind-turbines were measuredin a variety of meteorological conditions. Strain gauges were also located within the nacelles ofthe wind turbines and load cells within each of the three catenary mooring lines. In the lasttwo test phases alone, over 2000 hours of offshore data were recorded. This paper details anapproach to validating the HAWC2-WAMIT numerical model using this offshore data.

Comparisons are made in this paper of the numerical model to a theoretical floating turbinefor which published results are available, and to offshore data from the 4th operational stageof the P37. Initial results are given in this paper considering the motions of the platform andloads on a mooring chain, together with an outline of the next steps to take in the process ofvalidating the numerical code by comparison to the offshore data.

2 The Floating Power Plant Offshore Device, P37

The offshore data to which DTU model, HAWC2-WAMIT, will be compared was measuredon the 37 m wide hybrid wind- and wave-energy device, P37, developed by Floating Power

2

(a) (b)

Figure 1: Artists impression of Floating Power Plant’s P80 device

Plant A/S. This section offers a brief history to the development of the device followed by adescription of it.

2.1 Development Background

Floating Power Plant are the developers of a novel, floating, wave- and wind-energy hybriddevice. The design for the first commercial device, the P80, is currently being finalised. Inshort, the P80 will consist a semi-submersible platform which will be approximately shapedlike a cross, with a central hull, an 80 m wide perpendicular cross-bridge, and damping plateslocated far beneath the free surface. At the fore end of the hull, a mooring turret will belocated, about which the platform will be able to rotate freely in order to face the incomingwaves. Across the cross-bridge, four pitching wave energy converters will operate, each attachedto its own PTO house, which will be an oil hydraulic system housed within a dry ’engine room’.On top of the hull, a single wind turbine will be located. The unique, patented technologyprovides a stable platform for any offshore wind turbine, hence is not specific to any one windturbine manufacturer.

Since the initial conception of the invention, Floating Power Plant have performed hydro-dynamic tests on their device at 1:16 and 1:14.5 scale in wave flumes, 1:33, 1:9.5 and 1:50 scalein wave basins, and 1:50 scale in a combined wind and wave facility with the purpose of designdevelopment and optimisation. Special attention has been paid during the research and devel-opment to the patented wave energy technology, in which the unique geometry of the floatersinteract with the platform to maximise their performance, the stability of the platform and thePTO control system.

A key part of the development has been the design, build and operation of Floating PowerPlant’s 37 m wide test platform. This has been undergoing offshore testing for four completetest phases (totalling more than 2 years). The test platform provides electricity to the gridfrom both wind and wave energy, however its purpose is purely for research and development.

The first test phase focused on the platform and WECs and the installation of appropriatemeasurement sensors. The second test phase included the installation of the wind turbines.The third test phase focused on the installation of the PTO system, and, following wave flumetests on the WECs, the replacement of two of the WECs with different sizes more suited to theP37 wave conditions. The fourth test phase focused on testing different control strategies forthe PTO system.

3

The PTO for the WECs has undergone rigorous testing and optimisation through both dryand offshore testing. The key drivers for this optimisation are both increasing the power outputacross a wide range of wave conditions and smoothing the power output to meet grid demands.Through refining the PTO control strategy, the mean wave to wire efficiency (including all datafrom the entire test period, and all control strategies tested), was increased to over 30 % intest phase 4 [4]. Further details on the test undertaken up to February 2014 can be found inthe public report, [8]. Since this report, combined wind and wave tests were performed at 1:50scale in October 2014, and further dry testing of the PTO unit undertaken.

2.2 The P37 Offshore Test Device

The P37 is not a directly scaled version of the P80, but is instead specifically designed for aDanish offshore site in order to test and obtain operational data and knowledge on the keydesign components of the P80. Instead of four large wave energy converters, 10 smaller waveenergy converters (WECs) are operated across the cross-bridge. The eight inner WECs arelarger than the outer two WECs, in order to test the relationship between the WEC size andits motions in different waves. All floaters have the ability to be ballasted to tune their naturalperiod to the incident waves, however in the fourth test phase constant ballast was used inorder to compare different PTO control strategies.

A complex control algorithm, developed in partnership with Fritz Schur Energy A/S, isapplied to oil hydraulic PTO unit of the most starboard WEC. Throughout the test periods,the PTO algorithm has been developed and fine-tuned. In-between test periods, the PTOunit from this WEC was removed from the platform and dry-tested in a laboratory to furtherdevelop PTO control algorithms.

As with the P80, the motion of the WECs results in the passive vaning of the platform aboutits mooring turret so that the WECs face the incoming waves, with high efficiencies found whenthe WECs face within 30 degrees of the predominant incident wave direction. Also like the P80platform, the P37 is actively ballasted in order to maintain even keel in all conditions.

Unlike the P80, which has a single wind turbine of 2.3 to 5 MW, the P37 has three smallerturbines (Gaia 11 kW), one on either side and one on the aft of the hull. The wind turbinesare two bladed in a downwind configuration with teeter hinge, fixed pitch stall control and apassive yaw system that aligns the single rotor towards the incoming wind. The turbines wereselected in order to obtain the necessary data using off-the-shelf turbines with a hub heightappropriate to the platform size. The water depth of a site suitable for the P80 is 45 to 200m, so to adhere directly to Froude scaling, the P37 would need to be in a depth of 23 to 93 m,however the P37 site is only 7 m in depth. The catenary mooring system designed for the P37site therefore has different characteristics to the P80 system.

There have been innumerable lessons learnt during the operation of the P37 which will beinvaluable in the design and operation of the P80 device. Although there are key differencesbetween the P37 and P80 designs and sites, the vast amount of data gathered over the two yearoperational period, is used for testing and validating numerical models on all components ofthe device. Once validated, these same models can be used to enhance the design for the P80device.

This paper focuses on a model which combines the wind turbines, floating platform andmooring system, but neglects the motion of the WECs and the WEC PTO units.

4

(a) Aerial photograph taken from a helicopter on acalm day

(b) Photograph taken from P37 webcam during astorm

Figure 2: Photographs of Floating Power Plant’s 37 m wide offshore test platform, P37

3 The HAWC2-WAMIT Numerical Model

HAWC2 (Horizontal Axis Wind turbine simulation Code 2nd generation) is an aeroelastic codeintended for calculating the response of both floating and fixed foundation Horizontal AxisWind Turbines in the time domain.

The aerodynamic model within HAWC2-WAMIT is based on the Blade Element MomentumTheory, with modifications to account for wake expansion and swirl and extensions to accountfor dynamic inflow, skew inflow, shear effect on induction, the effect from large blade deflections,tip losses and two dynamic stall models (Stig ye and Modified Beddoes-leishmann models).

The flexible dynamic mooring line is model by a general cable element formulation which issimilar to e.g. ([1]). The model is able to handle mooring line systems that includes cable andbeam elements, as well as buoys and clump weights. Springs are included to account for bottomcontact but are only activated when there is a contact between sea floor and the line. Theformulation used allows for very large deformations of the mooring lines. Hydrostatic buoyancyand hydrodynamic drag forces are considered throughout the analyses. This formulation onlyholds for cables with uniform properties, so each section of different cable type (chain, syntheticrope etc.) is modelled as a separate body and connected by ball joint constraint in HAWC2smulti-body formulation. Point masses are implemented with both linear and quadric viscousdamping terms to model clump masses and drag buoys.

The mooring system model uses a special interface to HAWC2, which enables externallyformulated equations of motion to be solved together with the core equations of motion in theHAWC2 code ([5]). The external system interacts with the HAWC2 model through a set ofconstraint equations which describe how the external system degrees of freedom (DOFs) andthe HAWC2 DOFs are related. Both the EOM for the external system and the constraintequations are specified in an external DLL which is then called by the HAWC2 solver duringthe simulation.

The structural model includes flexible rotating bodies which include masses, and inertial,centrifugal and gyroscopic forces.

The model applied in this paper includes a standalone coupling to the wave analysis software,

5

WAMIT. The external DLL interface in HAWC2 is used to read the WAMIT output files directlyand input to a HAWC2-WAMIT module, ESYSWamit, in a similar way to the coupling withthe mooring module.

A more detailed description of the model is given by [3].

4 OC4 Comparison

The International Energy Agency (IEA) devised the OC4 project ([7]) under Wind Task 30in 2010 to investigate wind-turbine coupled simulations with both jacket structures (Phase I)and semi-submersible platforms (Phase II). The aim of the project was to perform code tocode comparisons of state-of-the-art simulation codes for offshore wind turbine modelling. Inphase II, 21 load cases were applied with varying levels of complexity and a variety of metoceanconditions. The cases are clearly defined, hence a small selection of them are applied here foran initial comparison of the FloVAWT and HAWC2-WAMIT codes.

The device to which these load cases are applied is a semi-submersible floating offshorewind system developed for the DeepCwind project. The platform is 20 m draft, with a tripodconfiguration with a distance of 50 m between each of the three circular legs. The turbine isthe NREL 5MW reference turbine which has a tower height of 77,6 m, attached to the platform10 m above the free-surface when at rest.

Out of the 21 load cases defined in the OC4 Phase II code comparison, only 3 are appliedas this comparison is not the key focus of this paper, Load Cases 1.3, 2.1 and 3.1. The resultsare discussed in the following sections.

(a) Surge (b) Heave (c) Pitch

Figure 3: OC4 load case 1.3a: HAWC2-WAMIT (black dashed line) overlaid onto Figure 4 from[7]

6

(a) Surge (b) Heave (c) Pitch

Figure 4: OC4 load case 1.3b: HAWC2-WAMIT (black dashed line) overlaid onto Figure 5from [7]

(a) Surge (b) Heave

(c) Pitch (d) Fairlead Tension 2

Figure 5: OC4 load case 2.1: HAWC2-WAMIT (black dashed line) overlaid onto Figure 6 from[7]

5 Implementation of P37 in HAWC2-WAMIT

HAWC2 (Horizontal Axis Wind turbine simulation Code 2nd generation) is an aeroelastic codeintended for calculating the response of both oating and xed foundation Horizontal Axis WindTurbines in the time domain.

The aerodynamic model within HAWC2-WAMIT is based on the Blade Element MomentumTheory, with modications to account for wake expansion and swirl and extensions to account

7

for dynamic inow, skew inow, shear eect on induction, the eect from large blade deections, tiplosses and two dynamic stall models (Stig Øye and Modied Beddoes-leishmann models).

Within the mooring module, an iterative solution of the quasi-static equilibrium betweenthe free anchor chain and the angle at the connection point of the chain and the platform isused to compute the bottom contact point. The horizontal and vertical components of the lineforce at the connection point are then determined from the weight of the oating chain and theangle of the chain at the connection point. A fully dynamic analysis is then performed on themooring line from the connection point to the bottom contact point. This is performed bydividing the line into sections, a uniform stiness, mass and hydrodynamic characteristic, witheach section then subdivided into 2-node elements on which the analysis is performed.

The structural model includes exible rotating bodies which include masses, and inertial,centrifugal and gyroscopic forces.

The model applied in this paper includes a standalone coupling to the wave analysis software,WAMIT. The external DLL interface in HAWC2 is used to read the WAMIT output les directlyand input to a HAWC2-WAMIT module, ESYSWamit. A more detailed description of themodel is given by [3].

5.1 Implementation Within the HAWC2-WAMIT Modules

The complex submerged structure of the P37 was simplified for implementation in WAMIT, byremoving smaller pipes and substructures. The simplified submerged structure was convertedto a system of panels, each with 3 or 4 vertices, for implementation in WAMIT using the 3DCAD software, Rhino, together with the mathematical software, MATLAB (see Figure 6).

(a) 3D CAD drawing (b) Panel model used within WAMIT

Figure 6: 3D CAD drawing of P37 and the corresponding simplified panel model of the sub-merged geometry implemented in WAMIT device

The P37 platform is actively ballasted to maintain close to even trim in different conditions.The exact ballasting recordings were not always recorded accurately and have not yet beencorrelated with the processed measurement data. Based on the design conditions the centre ofgravity has been calculated by naval architects, however in reality this will vary according tothe exact ballasting conditions.

As the initial measurement period selected is a period during which the wind turbines arenon-operational, the initial test case does not include any wind applied to the model. In future

8

iterations, the wind will be applied even with the wind turbines non-operational, as the dragfrom the wind may influence the platform behaviour.

The three spread catenary mooring system is implemented into the mooring module, con-nected to the mooring turret within the anchor buoy at the front of the hull. As the platformrotates about the mooring point, the motions are given with the origin at the intersection ofthe undisturbed free surface and the mooring turret.

5.2 Selecting Suitable Offshore Data Period

The P37 is not an exact scale of the commercial platform, P80, but is instead designed forobtaining suitable offshore measurements. Sensors are placed on all aspects of the device,including load links in each of the mooring lines, two inclinometers on the platform to determineplatform pitch and roll, gps to track platform direction, a sonic anemometer to measure windspeed and direction, the current direction and speed, and the mean wave statistics . Themeasurements are stored in 10 minute intervals, with most measurements recorded every 0.03seconds (35 Hz sampling frequency), but some, such as the wave statistics, recording only amean value for each 20 minute interval.

The platform contains 10 pitching wave absorbers, whose motion is closely integrated tothe hydraulic PTO system. The floaters are each partially enclosed by the platform, resultingin complex hydrodynamic properties. A separate numerical model is under development forthe floaters, but at this stage is not coupled to the HAWC2-WAMIT model. For this papertherefore, a measurement period where the floaters are fixed (vertically upwards) is required, toavoid the complex interactions of the waves, the floater motions, the PTO unit and the rest ofthe platform. The HAWC2-WAMIT model is theoretically able to model multiple wind turbinesin turbulent wind conditions, and include their influence on the motion of the floating platformtogether with the mooring forces. At this preliminary stage of comparison, a more simplifiedcase is required, however, in order to understand the comparisons between the simulations andoffshore measurements in detail. The measurement period is therefore also required to havethe wind turbines not in operation. Further simplifying the comparison case, the measurementperiod would ideally have a low current and wind speeds.

The third and fourth test phases contain around 2000 hours of data. As with all test data,however, there are often uncertainties in the recordings. A verification and filtering strategywas applied by DTU during post processing, comparing signals to each other to ensure theiraccuracy ([8]).

Using a this strategy, the measurement period selected as the most suitable period for thisfirst stage of simulation comparison is a one hour period starting at 11:40 on 5th December 2012.During this period, nine out of ten of the floaters were in the storm position (fixed verticallyupwards), the turbines were not operating and the current speed was low (0.4 m/sec). Theaverage significant wave height was, Hm0 = 0.61 m , the peak period was Tp = 3.4 seconds andthe misalignment between the waves and the platform was an average of 15 degrees. Althoughthe effects of wind will not be included in the simulations in this paper, the mean wind speedduring the selected measurement period was relatively high at 9.7 m/s, which is likely to haveinfluenced the platform even with the wind turbines non-operational.

9

5.3 Defining a Spectrum

The wave statistical measurements recorded at the P37 site are averaged over 10 minute in-tervals. As no time series data exists for waves at this site, it is impossible to know whichtheoretical spectrum would better represent the waves at the site. An irregular wave field canbe approximated by summing component regular waves with a range of phases, amplitudesand frequencies [6]. The distribution of energy between these component waves can be definedby a theoretical spectrum, with different distributions of energy resulting in the same overallstatistical wave properties. Common theoretical spectra include the Pierson Moskowitz andthe Jonswap spectra. The Pierson Moskowitz spectrum was originally developed based on mea-surement data from the North Atlantic (open sea) and the Jonswap from the north sea (limitedfetch). The two theoretical spectra are similar in definition, with the Jonswap including apeak enhancement factor, γ. A Jonswap spectrum with γ = 1 exactly represents the PiersonMoskowitz spectrum, however increasing the γ value, increases how pronounced the peak ofthe spectrum is, hence decreases the energy in the low and high frequency ranges. Where theexact specifications of the energy distribution are not known, a standard γ value of 3.3 can beapplied.

In this paper both the Pierson Moskowitz and Jonswap with γ = 3.3 will be applied. Whilstthe spectrum defines the distribution of energy between the component waves, it does not definethe phases of the waves, hence these must be defined randomly. Three ’seeds’ of each of thetwo spectra are therefore applied with different random phases in each seed.

6 Preliminary Results of Simulations

Comparing time series data from simulations and measurements is not straight forward. Thereare many statistical measures that can be used, as well as visual comparisons of the data. Table1 shows some basic statistics, which can be used as an initial indicator. Further statisticalcomparisons will however be included in future iterations.

10

Offshore JS1 JS2 JS3 Mean Js PM1 PM2 Mean PM

Mean value 0.52 -0.1 -0.1 -0.1 -0.10 -0.1 -0.1 -0.10

Mean Amplitude 0.50 0.45 0.45 0.4 0.43 0.85 0.84 0.85

Mean of highest1/3 of amplitudes 0.69 0.57 0.58 0.52 0.56 1.06 1.04 1.05

(a) Pitch (degrees)

Offshore JS1 JS2 JS3 Mean Js PM1 PM2 Mean PM

Mean value -0.28 0 0 0 0.00 0.01 0.01 0.01

Mean Amplitude 0.29 0.1 0.1 0.1 0.10 0.18 0.15 0.17

Mean of highest1/3 of amplitudes 0.41 0.12 0.13 0.12 0.12 0.23 0.19 0.21

(b) Roll (degrees)

Offshore JS1 JS2 JS3 Mean Js PM1 PM2 Mean PM

Mean value 2.98 3.04 3.04 3.04 3.04 3.03 3.03 3.03Mean Amplitude 0.25 0.21 0.20 0.19 0.20 0.20 0.17 0.19Mean of highest

1/3 of amplitudes 0.36 0.27 0.28 0.25 0.27 0.28 0.25 0.27

(c) Mooring Load on mooring line 1 (Tonnes)

Table 1: Initial statistical comparison between offshore data from P37 (averaged over one hour)and HAWC2-WAMIT results with the Jonswap spectrum (seeds titled JS1, JS2 and JS3) andPierson Moskowitz spectrum (seeds titled PM1 and PM2)

6.0.1 Pierson Moskowitz Compared to Jonswap

The mean amplitudes of the pitch and roll motions are approximately doubled in the PiersonMoskowitz spectra compared to the Jonswap spectra. This is due to the shift in the wavespectrum towards lower frequencies, resulting in a wider PSD peak in the pitch motion, as seenin Figures 7 and 8.

11

(a) Time series of free surface elevation (m) (b) PSD of free surface elevation, all frequencies

(c) PSD of free surface elevation, low frequencies

Figure 7: Time series and spectra of the free surface elevation Jonswap and Pierson Moskowitzspectra, JS1 = red and PM1 = black

(a) Time series of platform pitch (deg) (b) PSD of platform pitch

Figure 8: Time series and spectra of the free surface elevation and the pitch platform motionwith the Jonswap and Pierson Moskowitz spectra, JS1 = red and PM1 = black, and measureddata = blue

12

The amplitudes of pitch motion are too small in the simulations with the Jonswap spectraand too large in Pierson Moskowitz spectra compared to the offshore measurements. Thisindicates that neither theoretical spectra accurately models the P37 site, and perhaps a thirdspectrum could give better results.

6.1 Mooring Loads on Mooring Line One

The mooring loads are similar in the simulated result to the measured data, however theamplitudes are slightly too small. The high frequency behaviour of the mooring lines is notcaptured in the simulations. This needs further investigation, together with application to othermeasurement periods where offshore data is available for more than one mooring line.

6.2 Limitations

In the selected offshore measurement period, the waves come from an average of 15 deg misalign-ment, hence platform yaws about its mooring turret at the front of its hull. The hydrodynamicinput to HAWC2-WAMIT is from WAMIT with the loads given about the centre of flotation.As the platform moves about its mooring turret, the input from WAMIT is not modified toaccount for the new position.

As the P37 is a test platform, designed for research and development purposes, many mea-surements were taken, but also many things were changed throughout its test periods. Onesuch variation is the ballasting of the platform, which varies its centre of gravity. Both a centreof gravity calculated for a single ballasting condition, and second centre of gravity calculated inattempt to match the ballasting situation offshore applied in this paper. It is shown the choiceof centre of gravity impacts the results. Further work therefore needs to be performed, therefore,to cross-check all records to accurately determine the centre of gravity for each measurementperiod in order to perform future model validations.

In order to validate the numerical model, the meteorological conditions seen offshore needto be replicated in the model. The statistical wave data recorded at the P37 offshore test sitedoes not contain enough information to accurately repeat the conditions within the model. Twotheoretical spectra are applied in this paper, however they result in significantly different resultsin both the platform motions and mooring loads, due to their difference in energy distributionat high and low frequencies. Further investigation is therefore needed to determine the mostappropriate theoretical spectrum to replicate the wave conditions seen by the P37 device.

The method by which the hydrodynamic module is currently integrated to the mooring andaerodynamic modules does not accurately model yaw motion. For this purpose, the comparisonsin this paper have focused on the pitch and roll motion of the platform and the mooring loads.Once corrected, further validation can be made to offshore data when considering the yawmotion. It is difficult to draw conclusions on the mooring line simulation based on only onemeasurement period and one mooring line. Once the potential to model the yawing motionabout the mooring turret is included, further validation cases must also be identified in whichdata is available from multiple mooring lines.

7 Conclusion

One of the key challenges facing the development of floating wind devices is the development ofreliable numerical models to predict the loads and motions on the devices under different sea

13

and wind conditions. HAWC2-WAMIT is a code under development by DTU which modelscomplex wind conditions, wind turbines loads and motions combined with the loads and motionsof a floating foundation and the mooring system. Floating Power Plant is the developer of afloating hybrid wind and wave energy device, and has 2 years of offshore operational data. Inthis paper, the challenges of applying the numerical model to the offshore data are exploredwith a view to validating the numerical model. The comparisons performed in this paper arethe first iteration, with a view to identifying the areas which need clarifying to perform detailedvalidations in the future.

Despite a significant amount of data, a large amount of post-processing and quality checkingmust be performed on the data to identify suitable periods. As well as good quality data, it isalso crucial to identify areas where the complexity is minimised for early stage model validations.A test period with the wind turbines and wave energy devices non-operational is applied as apreliminary simple test case.

Based on the comparisons made in this paper, three key areas have been identified forimprovement before the next stage of model validation. Firstly, it is deemed necessary tocarefully cross-reference all offshore records and measurements to establish the exact centre ofgravity of the platform at all times. Secondly, studies need to be performed on the test siteto identify the most suitable theoretical wave spectra to accurately represent the wave climateassociated with the measurements. Thirdly, the method in which the hydrodynamic loadsand forces are integrated into the model needs to be modified to account for platform yawingmotions resulting from misalignment of the platform with the waves.

Further validation stages will include identifying offshore testing periods with high qualitydata, with increasing complexity (such as with wind turbines operational).

A separate model is also under development within Floating Power Plant for the combinationof a single wave energy converter with their hydraulic PTO system. This is currently undergoingsimilar validation studies. Once validated separately, this model will be integrated into theHAWC2-WAMIT model to give a fully integrated model of the P37 platform with both windturbines and wave energy convertors in operation

8 Acknowledgements

This article and background work is done with funding from The European Regional Devel-opment Fund. Floating Power Plant would also like to thank the European Union SeventhFramework Programme H2Ocean project for inclusion in the project and for their supportthroughout the project.

References

[1] P. B. Bergan, E. Mollestad, and N. Sandsmark. Non-linear static and dynamic responseaanalysis for floating offshore structures. Engineering Computations, 1985.

[2] Arapogianni. A et al. Deep water the next step in offshore wind energy. Technical report,EWEA, July 2013.

14

[3] Anders M. Hansen. Implementation of hawc2 interface to wamit. Technical report, TechnicalUniversity of Denmark, Department of Wind Energy, 2014.

[4] J Hansen. P37 hydraulic pto test phase 4 summary. Internal Report.

[5] B. S. Kallesøe and A. H. Hansen. Dynamic mooring line modelling in hydro-aero-elasticwind turbine simulations. In Proceedings of the Twenty-first (2011) International Offshoreand Polar Engineering Conference, 2011.

[6] A.F. Molland. The Maritime Engineering Reference Book: A Guide to Ship Design, Con-struction and Operation. A Guide to Ship Design, Construction and Operation Series.Elsevier Science, 2011.

[7] A. et al Robertson. Offshore code comparison collaboration continuation within iea windtask 30: Phase ii results regarding a floating semisubmersible wind system. In Proceedings ofthe ASME 2014 33rd International Conference on Ocean, Offshore and Arctic Engineering,2014.

[8] A. et al Yde. Experimental and theoretical analysis of a combined floating wave and wind en-ergy conversion platform experimental and theoretical analysis of a combined floating waveand wind energy conversion platform. Technical report, Technical University of Denmark,Department of Wind Energy, 2014.

15